Smartling's MCP (Model Context Protocol) server allows you to integrate Smartling with AI assistants and LLM tools (like Claude Code, Visual Studio Code, Cursor, and more), enabling access to Smartling's translation capabilities directly within your developer environment or AI chat tool interface.
Join our Smartling Community to view a video walk-through of this feature. You can also find Smartling's MCP server in the Anthropic MCP Registry.
What is MCP?
The Model Context Protocol (MCP) is an open standard developed by Anthropic that provides a standardized way for AI applications to communicate with external systems. You can think of MCP as a universal connector that allows AI assistants to securely access and interact with various resources such as databases, APIs, and other external platforms.
Smartling's MCP server acts as a bridge between your AI tools and Smartling's translation platform. Once you connect the MCP server to your AI tool, you can use the prompt interface to initiate actions using the tools made available by the Smartling MCP server.
What can I use it for?
Once you connect the MCP server to your AI tool, you can:
- Initiate actions in Smartling directly from the tool's chat interface
- Retrieve translation assets and information directly from Smartling
- Integrate Smartling’s translation capabilities into your existing AI-powered processes
For example, you can instruct the AI tool to translate your content without the need for any complex configurations. Instead of returning a generic translation that is not tailored to your industry or preferences, you can easily integrate your Smartling assets using a simple prompt. This produces a more customized translation output that adheres to your preferred terminology and style. Similarly, you can create prompts to look up existing translations or information from your Smartling account.
This means that instead of logging into the Smartling dashboard, you can simply access your AI tool and use the prompting interface to make use of Smartling's MCP tools.
This list will continue to grow as Smartling releases new MCP tools for additional AI assistant use cases.
Smartling MCP tools
Smartling provides various MCP tools that act as building blocks. These tools allow the connected LLM to access information about the various translation resources across your entire Smartling account. For example, Smartling's MCP tools allow a connected LLM to access your linguistic assets (such as your glossary, style guide, and translation memory) or translation data (like a list of all projects or jobs).
Your AI tool can read from these tools and call on them as needed, based on your prompts, to achieve the desired results. Once the Smartling MCP server is installed, all available MCP tools will appear in your AI tool, along with a complete description of each tool.
Example:
The information below provides an overview of the available MCP tools.
Text translation
- Translate individual sentences, paragraphs, or text blocks
- Use your preferred NMT provider or LLM through Smartling's MT API
- Apply linguistic assets (glossaries, style guides, and translation memory) to enhance translations
- Generate customized translation output that adheres to your preferred terminology and style
File translation*
*File translation support is currently limited due to MCP security restrictions that prevent uploading files to remote servers. Limited support exists for small text-formatted files with size restrictions. Smartling continues to monitor MCP policy changes and will enhance file translation capabilities as new options become available.
Access linguistic assets and account information
For example:
- Retrieve and list glossaries, style guides, and linguistic packages
- Access translation memory for existing translations
- Search and match terms within specific glossaries
- Get information on Smartling projects and jobs
- Get a list of users registered under your Smartling account
Initiate actions in Smartling
For example:
- Add or remove string tags
- Authorize translation jobs
- Add strings to an existing translation job
Supported AI tools
MCP is a relatively new technology and is not yet universally supported across all AI tools. However, it is supported in popular AI-powered code editors such as Claude Code, Visual Studio Code, and Cursor, which can be used to access the Smartling MCP server.
We recommend checking the documentation for your preferred tool to confirm whether it supports connecting to an MCP server.
Setup instructions
An MCP token is required to connect your AI tool to Smartling’s MCP server. To create an authentication token, you must have an Account Owner role in Smartling. If you do not have this role, you will need an Account Owner to create the token and share it with you.
Step 1: Configure your translation settings in Smartling
Linguistic assets
To produce the desired translation output, your linguistic assets in Smartling should be up to date and of good quality. We recommend reviewing these assets before starting to translate.
Translation provider
Unless otherwise specified in your prompt, the LLM will use the translation provider configured in your MT API settings.
From the top navigation of your Smartling dashboard, go to AI Hub > Instant MT > select the MT API tab.
Here, you can select the MT profile that should be used for translation requests made through the MCP server. Be sure that the toggle is switched On.
Please note that Glossary Term Insertion will need to be requested specifically in your prompt, even if it is enabled on this page.
Step 2: Connect the Smartling MCP server to your AI assistant
MCP Host
To establish a connection with Smartling, an MCP host is required. Popular AI assistants and AI-powered code editors like Claude Code, VS Code, and Cursor support MCP and can be used to access the Smartling MCP server.
Generate an authentication token
To allow the MCP host to access Smartling's tools, you need to generate an authentication token. Tokens can be created and managed in the Smartling dashboard by users with an Account Owner role:
- Go to Account Settings > Integrations & Developer Tools, and click MCP Tokens from the menu on the left.
- Click Create Token.
- Enter a token name that is easily identifiable by your team.
- To confirm, click Create Token.
Once the token has been generated, keep this page open. In the next step, you will need to copy and paste the token into the configuration file for the connecting application.
Configure the Smartling MCP server
The exact configuration process depends on the MCP host application you use. The instructions below are specific to Visual Studio Code using GitHub Copilot.
- Create a new folder in your file system
- Open the folder in VS Code and add a subfolder called
.vscode - Within the
.vscodesubfolder, create a configuration file namedmcp.json -
Copy the JSON configuration into the
mcp.jsonfile, replacing "YOUR_TOKEN_HERE" with your actual MCP token. Smartling provides configuration examples for several tools, including VS Code. Below is an example configuration.{ "servers": { "smartling": { "type": "http", "url": "https://mcp.smartling.com/mcp", "headers": { "X-SL-Authentication": "<YOUR_TOKEN_HERE>" } } } } - Save the file and start the server.
Your AI tool is now connected to Smartling's MCP server and can use the available tools.
Viewing and managing available MCP tools
Once the connection has been successfully established, you can view and access all available Smartling tools. In VS Code, this can be done by clicking the "Configure Tools" icon from the chat window.
By default, all Smartling MCP tools are accessible. They allow the AI tool to access various account-wide resources and information from Smartling, such as all of your Smartling glossaries, projects, and style guides. If you would like to disable any of the tools, you can deselect them and save the configuration.
Step 3: Create a prompt
You are now ready to start prompting. Once the Smartling MCP server has been configured, all requests can be initiated directly from the chat interface of your AI tool using natural language prompts. MCP provides a highly flexible framework allowing you to use and combine all available tools to achieve the desired result.
For example, you can ask the LLM to translate a text snippet into a specific target language using Smartling for machine translation.
In your prompt, you can specify additional resources or linguistic assets, such as terms from your Smartling glossary. Be sure to mention that you want to use Smartling for translation.
Example:
“Please translate this text into de-DE using Smartling machine translation. Use terms from the Marketing Glossary.”
If you don't specify in your prompt that you want to use Smartling, the LLM will typically return a generic translation instead of using Smartling's MT API.
The LLM will then use all required tools for your request and return a translation. If clarifications are needed, the LLM may ask you to confirm the Smartling resources that should be used to process your request. The LLM will identify individual tools—click Continue to run each tool.
Ultimately, the LLM decides which tools to use when interpreting your prompt. For example, if you want terms from your Smartling glossary applied, specify this explicitly in the prompt. We recommend using prompts that are as specific as possible to achieve the desired outcome. Smartling cannot guarantee specific translation results.
Your prompt can also point to other Smartling tools to retrieve existing translation data from your Smartling account. For example, you can fetch the existing translation for a string stored in a specific translation memory or project.
Important considerations
Smartling's MCP server only supports instant text translation via MT or LLM. Uploading files to a Smartling project with string parsing and translation workflows is not currently supported.
When using this tool, no files or text are uploaded to a Smartling project and translations are not saved to your translation memory. Content can only be translated using MT or LLM translation without a human in the loop. The translations do not pass through a translation workflow.
Any content that is requested is not stored or ingested into Smartling.
MCP is an innovative and fast-growing technology. The list of available Smartling MCP tools will continue to evolve to support a wide number of different use cases. If you are interested in using the Smartling MCP server, please feel free to contact us. Our team will be happy to assist you and evaluate how to best achieve your goals.